Grok says....

....I should retire and replace my 4 x 3TB WD Green, 2 year old power on, disk drives being used as backup in a USB dock. Even though SMART is reporting good health. The reason for replacement according to Grok and Copilot is a number of uncorrectable sectors (under 100 raw data) and SMART is 200 for that attribute with a threshold of 140.

Currently running write and verify tests on the disks as suggested by Grok. What do you reckon?

In reply to a post by Banger:

....I should retire and replace my 4 x 3TB WD Green, 2 year old power on, disk drives being used as backup in a USB dock. Even though SMART is reporting good health. The reason for replacement according to Grok and Copilot is a number of uncorrectable sectors (under 100 raw data) and SMART is 200 for that attribute with a threshold of 140.

Currently running write and verify tests on the disks as suggested by Grok. What do you reckon?

I reckon that cloning disk images is not part of a viable backup strategy with such hardware

So you think the HDDs are pre-fail too despite SMART being good?

No, I think that they have failed already if your backup strategy is disk cloning. I would have a different view for file backup and incremental file backup.

It is using Macrium file backup once a week with differential updates and once a month a full clone. All my critical documents, photos and music is store on Onedrive (cloud) and that is my offsite.

Edited by Banger (Mon 17-Feb-25 23:06:03)

I have lost track of the previous thread, but I thought you said your highest level backup under your new plan was going to be a clone.

Clone once a month as with file backup every week with having Linux and Windows on the same disk file backup isn't enough to restore partition file and boot partitions as well as files.

So in essence I can restore up to date files then the partition and boot files from the clone - its selective. It's a hybrid approach.

Anyway the question was about the HDDs.

I think you are relying on advice from a source that is definitely Artificial but not Intelligent.

The absolute number of bad sectors on a hard disk is of little importance, especially with large drives. Whether that number is increasing is important.

My thoughts exactly. I think these errors are historical as there is zero pending sectors and zero reallocated sectors. Found a nifty piece of software called Victoria which scans the HD for bad sectors and depending upon mode tells the drive to remap any bad sectors or refresh them at software level. Scan is at 75% on a 2.5TB drive and there aren't even any slow to access blocks eg over 3s access time which is usually a sign of a block going bad.

In reply to a post by Banger:

Anyway the question was about the HDDs.

Obviously. And whether you are cloning or backing up files is relevant to that.

In reply to a post by Banger:

Clone once a month as with file backup every week with having Linux and Windows on the same disk file backup isn't enough to restore partition file and boot partitions as well as files.

So in essence I can restore up to date files then the partition and boot files from the clone - its selective. It's a hybrid approach.

If you have that much of a disaster, you can reinstall and then restore the files. Hopefully you have your data on its own partition.